MSPhysics 1.0.3 (16 October 2017)

[Anton_S]

Hi, Ecati,

@ecati said:

I do not want to bother you with such simple questions. But I have not found a tutorial that will answer my questions. If I'm not wrong, you're preparing a tutorial, are not you?

I first need release version 1.0.0. Then I will transition to making tutorials. The tutorials is a bold part on my TODO list.

@ecati said:

Now, how can I bring the green piston into its original position with a spring after it is pulled out?

You can't have the green thing be connected to a piston and the spring at the same time. You can simply tell the piston to move back to its original position by dragging its associated slider controller to its original position. You can as well change the controller of the piston joint to do it with a click of a key, such as key('f') * 5, so that it moves 5 preset units (millemeters in your case) when key F is held down and back to initial location when the key is released. You can add a negative sign to 5 or in front of the controller for it to move in the opposite direction.

@ecati said:

Collision problem.
When the first part is moved, it enters into the second part and the third part escapes from the center. How is this prevented?

There must be something in the way. Here is a few potential causes to this:

1. Part2 is connected to multiple joints, such as the hinge and the piston. It must be connected to just the hinge.
2. Either part2 is pushed out by part3 (if the hinge is not inside part3 or part3 state is not set to non-collidable). Either move hinge inside part3 and set it static or set part3 non-collidable so it doesn't push out part2.
3. Another collision outside the shown view is stopping part2.
   There could be many reasons. I won't know until I inspect the actual model myself, but the actual cause may very likely denote one of the provided potential causes above.

Regards,
Anton

[Anton_S]

Hello, Martin,

@ballbearing said:

I got the worm gear to do it's work quite nicely (MSPhysics & Sketchy). Just making the collision mesh groups is time consuming. I wonder if there exist possibility for something like in Blender & mesh collision option? The collision frame is not so super exact...

Those images show it in SketchyPhysics, but it will work the same in MSPhysics.
So, another way of doing it (in MSPhysics) is by using geared joints:

1. Set both gears to non-collidable and remove all original joints connecting the gears.
2. Add a Piston joint at the location and in the direction of the wormgear and connect it to the wormgear (using the Joint Connection Tool).
3. Add a Hinge joint at the center and in the normal direction of the gear and connect it to the gear.
4. Using the Joint Connection Tool, connect Piston joint to the Hinge joint to create a geared joint.
5. Adjust the gear ratio of the piston-to-hinge gear using the MSPhysics UI.
   Make sure you're using MSPhysics toolbars for doing all that...

Regards,
Anton

[ecati]

Where do we download MSPhysics 1.0.0?

[TIG]

@ecati said:

Where do we download MSPhysics 1.0.0?
You could wait [joke ] ?
The current version is 0.9.9 - and it's latest version is always available from the SketchUcation PluginStore:
http://sketchucation.com/pluginstore?pln=MSPhysics
Like most Plugins/Extensions...

[faust07]

Hi, ecati,
Sorry, I used a test version for the connect example file.
Anton is working on release version 1.0.0 - which soon comes out.

[SynSuka3D]

Bonjour,
Je voudrais lire la vitesse D'un corps sur l'axe Y interne,no axe Y mondiale !!!
J'ai essayé get_velocity...
set_var("speed_Y",????)

Hello,
I would like to read the speed of a body on the internal Y-axis, no Y-axis world!
I tried: get_velocity...
set_var("speed_Y",????)

[Anton_S]

Hello, Nicolas,

To get local velocity, along the body's Y-axis, you got to transform the global velocity relative to the body first, then get the Y component:

loc_vel = this.get_velocity.transform(this.get_normal_matrix.inverse)
vy = loc_vel.y

Regards,
Anton

[Anton_S]

Faust, I will ckeck out that cloth thingy soon.

[faust07]

Hi, Anton,
first test of a (cloth-) mesh with your MSPhysics API self connect script.
Unfortunately, each node joins only one rod. What am I doing wrong?

---

Connect_test_01.skp

---

[Anton_S]

Okay, faust.

Your script for connecting bodies attempts to connnect multiple bodies to the same joint. That's how it is in the user interface of MSPhysics, but behind the scenes a new joint instance is added for each constraint. So here is the script to have:

onTick {
  if frame == 1
    this.touching_bodies(true).each { |body|
      next if body.group.name != 'capsule'
      joint = MSPhysics;;BallAndSocket.new(this.world, this, this.get_matrix, this.group)
      joint.connect(body)
    }
  end
}

Paste it into one knot, and assign it to all knots, using the Assign Script to All With Name button.

That said, this is a very inefficient way of creating cloth, not to mention that you also have to create faces and then reposition vertices and map the UVs of textures (if any).

I already got my own, more efficient, version of cloth working, like 3 years ago. But it's not perfect. Here is the demo updated and refined for the current version of MSPhysics:

[cristyan]

Hello ! There is a script or another way that maintains the same distance between objects attached on CurvySlider , as a rollercoaster ? Thank you !

[faust07]

Thanks, Anton,
Your code works perfectly. For genuine cloth simulation this is, of course, too cumbersome. For me, it is a good practice to learn a little bit of good script again. In addition, the example shows the almost unlimited possibilities of MSPhysics.

---

---

[Anton_S]

Oops, the get_normal_matrix should be normal_matrix

Also, to quote other people, use the double-quote button or put them in quote tags surrounded with square brackets. And posting the whole video on the error isn't necessary either. Just post a message that is reported.

[SynSuka3D]

==============================

To get local velocity, along the body's Y-axis, you got to transform the global velocity relative to the body first, then get the Y component:
CODE: SELECT ALL
loc_vel = this.get_velocity.transform(this.get_normal_matrix.inverse)
vy = loc_vel.y

Regards,
Anton

===================
problem!!!!!!

[ecati]

This gears is not spinning properly. What would be the reason?

---

---

dişli.skp

[ecati]

Hi, Anton
Will you take a look at this file at a time when you are not busy?

---

makine.skp

[Anton_S]

Hello, Ecati,

For the first one, try setting material thickness of simulation tab, in the MSPhysics UI, to zero. It will improve the behavior by a bit. But because the gears are too small, the collision intersections aren't quite proper. You could increase world scale, say to 40, which should result in an improved behavior, but it will unfortunately result in a crash. I will try to fix that scaling bug before releasing 1.0.0.

For the second one, you can increase update rate, to say 4, and/or reduce update timestep, to say 1/120. The smaller the update timestep is, the more accurate the simulation is. The smaller update timestep should prevent collisions from deteriorating. Both of the settings can be adjusted through the Simulation tab of the MSPhysics UI dialog. After you do these, increasing the motor speed should have things running properly.

Nice examples by the way. I'm stunned with your improved MSPhysics skills.

Regards,
Anton

[ecati]

Is it possible to do servo controls via the msphysics API without writing a script?

[Anton_S]

It depends on how complex you want the controller to be. The simplest one is the slider controller. You can generate it by pressing the generate slider button in joint UI. Others are controlling with key, like key('f') as noted in some prior post. You can also write a function of time dirctly into the cpntroller, like: t = world.time; cos(t)*5 There are many controllers you can write without relying on the MSPhysics scripting API.

[Wolfy_71]

I'm really enjoying MSPhysics plug-in and have been following Sam Ketner's very insightful tutorials on Youtube. I'm wondering if there is a manual easily available, so I can bone up on some more stuff. More pressingly, I'm wondering how to attach one hinge with a bar (which appears to be in a fixed point) on a bar attached to another hinge... so that it is freely movable. I'm trying to draw an umbrella-like hinge at the moment and I'm stuck.